HAL Id: hal-01721678 https://hal.archives-ouvertes.fr/hal-01721678 Submitted on 2 Mar 2018 HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. Simulations of an aircraft with constant and pulsed blowing flow control at the engine/wing junction David Hue, Christophe François, Julien Dandois, Anna Gebhardt To cite this version: David Hue, Christophe François, Julien Dandois, Anna Gebhardt. Simulations of an aircraft with con- stant and pulsed blowing flow control at the engine/wing junction. Aerospace Science and Technology, Elsevier, 2017, 69, pp.659-673. 10.1016/j.ast.2017.07.031. hal-01721678
28
Embed
Simulations of an aircraft with constant and pulsed ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
HAL Id: hal-01721678https://hal.archives-ouvertes.fr/hal-01721678
Submitted on 2 Mar 2018
HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.
Simulations of an aircraft with constant and pulsedblowing flow control at the engine/wing junctionDavid Hue, Christophe François, Julien Dandois, Anna Gebhardt
To cite this version:David Hue, Christophe François, Julien Dandois, Anna Gebhardt. Simulations of an aircraft with con-stant and pulsed blowing flow control at the engine/wing junction. Aerospace Science and Technology,Elsevier, 2017, 69, pp.659-673. 10.1016/j.ast.2017.07.031. hal-01721678
Figure 17: Refinement and extension of the grid surrounding the slot.
Figure 18: The 14-slot system and its implementation on the aircraft.
16
Some additional details are given here concerning the last AFC device (n°6 and 7). It is composed of 14 slots of
70 x 6 mm with inter-spaces of 20 mm. It represents a length of 1250 mm in the spanwise direction. For the final
case (n°7), the 14-slot device is used as a pulsed jet system. The blowing frequency is 60 Hz (it has been determined
by other partner studies and following manufacturer constraints), the signal being rectangular, and the duty cycle is
0.5. Moreover, the slots blow alternatively: there is a phase shift of π between consecutive slots. The effect can
therefore be compared to the one of a sweeping jet with two output orifices. Figure 19 represents the blowing
velocities over a period at the exit of slots 13 and 14 (outboard side).
Figure 19: Blowing velocities of slots 13 and 14 for the device n°7.
VI. RANS simulations with constant blowing flow control
Only the AFC systems n°1 to 6 in Table 1 are addressed in this section. Besides, only ONERA performed AFC
computations on the considered configuration. As a consequence, all the results presented below are obtained with
the elsA solver. The aerodynamic conditions remain the same as in section IV.
Concerning the performance assessment of the different AFC systems and more specifically the global lift
coefficients, the results are shown in Figure 20. It can be noticed that some points are referenced as URANS, they
shall not be considered until the next section, they have been added only in order not to duplicate this figure.
The baseline configuration presented in Figure 10 is given here for comparison purposes. Also, red points are
visible. They correspond to computations of the baseline but with the surrounding slot grid added. It means that
there is no slot (no hole in the wing surface) but only the grid that will host the AFC system (see Figure 15). This is
only a check of a potential grid effect. As it can be noticed, the results that were obtained show no grid effect over
the whole polar.
17
The different AFC systems are presented according to their values, the other characteristics can be read in
Table 1. The first AFC device is a continuous slot of 2 mm with a low close to 0.09%. This setup is not powerful
enough to obtain a significant gain compared to the baseline configuration. However, the lift levels that are recorded
after stall are higher with AFC on, even with this device: at 17°, the lift coefficient is increased by +4%.
The second AFC system is still a 2 mm continuous slot but with a greater blowing velocity which leads to a
value of 0.15%. It can be seen that some lift gains are obtained this time. The CLmax is slightly increased (+0.025)
and it is achieved for an angle of attack of 16° instead of 15. Moreover, after stall, once the flow separation is
massively developed, the lift level remains relatively strong compared to the baseline curve (+ 5% at 17°).
Figure 20: CL(AoA); RANS and URANS AFC results; elsA.
Then, two identical systems are presented: they exhibit 7 slots 6 mm wide with a blowing velocity of 250 m/s,
which also corresponds to a value of 0.15%. They allow possible grid effects between the initial and refined slot
grids to be investigated, as explained above. It can be observed that the red and light green lines are almost
superimposed which is comforting about the initial slot grid refinement visible in Figure 15 and Figure 16.
Concerning this device in itself, it can be noticed that in pre-stall conditions, it is more efficient than the continuous
slot exhibiting the same value: a greater CLmax is achieved at 16° and it does not provoke deterioration of the
baseline polar at 12 or 14°. These aspects underline the potential of segmented slots. On the other hand, the post-
stall behavior is similar.
The 14-slot system, to be considered here in its constant blowing version (all the slots are blowing continuously),
exhibits a relatively high value of 0.21% which allows an overall better gain than the previous devices, even if
18
the CLmax coefficient is a bit lower. It is really interesting to observe that this system in particular seems to produce a
fairly smooth stall which might be valuable for aircraft handling qualities.
Finally, the AFC device with the highest value evaluated (0.26%) is a continuous 6 mm slot which requires a
mass flow rate of more than 2.5 kg/s which is probably too demanding in terms of bleed-air requirements. With this
system, the CLmax is significantly increased (+0.085) and stall is delayed of about 2°. This represents a substantial
3.5% gain in CLmax compared to the baseline configuration and the lift levels after stall are strongly improved.
Figure 21 is focused on the global drag coefficient. The impacts of AFC systems on drag are analyzed even if
this coefficient is not as important as lift during the landing phases. To start with orders of magnitude, it can be said
that both structured and unstructured (not shown here) computations of the baseline configuration produce
equivalent levels: roughly between 2000 and 5000 drag counts (one drag count stands for 1.10-4) for an angle of
attack increase from 12 to 19°. Concerning the baseline curve, it can be noticed that the massive flow separation
development generates a clearly visible and sudden drag rise that is delayed or even smoothed with AFC systems on.
However, it can also be observed that the blowing AFC devices produce a drag penalty before stall compared to the
baseline (but only of a few percent which might be considered as negligible for landing approaches).
Figure 21: CD(AoA); RANS AFC results; elsA.
Some local analyses will be given in the following paragraphs to explain the global coefficient evolutions. First,
the skin friction coefficient and friction lines are analyzed in Figure 22. Four cases are compared at 14, 16, and 17°,
angles of attack corresponding to pre-stall, stall, and post-stall conditions. These four cases are the baseline
configuration first and the AFC systems n°2 (2 mm continuous slot), n°4 (6 mm 7 slots), and n°6 (6 mm 14 slots)
defined in Table 1. As it can be observed in Figure 22, for pre-stall conditions at 14°, the flow is attached for all the
19
configurations (nevertheless a tiny beginning of flow separation can be seen at the wing trailing edge of the
baseline). Despite the flow is attached for the four cases, even at this relatively low angle of attack, some visible
differences can be noticed. Indeed, for the baseline (i.e. no flow control), some of the friction lines which start right
after the pylon tend to spread over a large part of the span, announcing a zone of weak velocities and then the
development of a trailing edge flow separation. On the other hand, for the configurations with flow control on, it can
be observed that the more the AFC system is efficient, the more the friction lines coming from the pylon remain
grouped. Then, at 16°, the difference for the baseline configuration is obvious: a massive flow separation has
appeared. In contrast, all the cases with AFC systems active still exhibit almost fully attached flow. However, some
discrepancies are noticeable between the different devices. The continuous slot and the 7-slot systems show flow
features that are similar to what was observed with the baseline at 14°. They both have friction lines that spread
downstream of the pylon and flow separations that start to appear at main wing trailing edge. Finally, at 17°, the four
cases presented here exhibit quite massive flow separation. Nevertheless, the fact that the 14-slot device produces a
relatively smooth stall can be understood here since its flow separation remains particularly limited compared to the
other ones. Besides, the last line of Figure 22 allows to explain why the lift levels after stall with AFC working are
much higher that the ones of the baseline configuration. Even if at a given angle of attack, the AFC systems do not
prevent the flow separation development that causes the stall, they are still able to contain it.
Figure 22: Cfx distribution and friction lines at 14, 16 and 17° (lines) for baseline and devices n°2, 4, and 6 (columns).
20
Still focusing on local insight, Figure 23 shows the flow patterns close to the flow control actuators. It exhibits
Mach number values in the slice defined in Figure 15 as well as through iso-surfaces that describe the flow injection
coming from the slots. The three devices presented are the same: n°2, 4, and 6. First, it can be observed that each
AFC system, by its blowing action, re-creates the acceleration normally due to the slat/wing slot (visible here on the
inboard side). Moreover, it can be noticed that the 6 mm slot devices (n°4 and 6), which have greater mass flow
rates, allow further propagation of the blowing velocity than n°2. In all cases, this propagation is not constant over
the AFC system span. It seems that the global flow (vortices coming from the inboard slat cutout and pylon in
particular) participates to the injection propagation downstream of the actuators. The Mach number fields in the
considered slice highlight how the injected flow spreads over the wing for each system. It is reminded that the mass
flow rate per slot of the 14-slot device (n°6) is lower than the one of the 7-slot device (n°4). This explains why the
injection of the third image of first line is not as strong as the one of the second image.
Figure 23: Slices and iso-surfaces (value=0.45) of Mach number showing the AFC injection at 16° for devices
n°2, 4, and 6.
The results presented in this section are consistent: as expected, the gain that can be obtained via the AFC system
is directly related to the value which represents the force given to the fluid. In these constant blowing conditions,
it seems that all the AFC systems evaluated allow significant lift level gains after stall; this is true even for devices
with values too low to improve the CLmax. The segmentation of the blowing slot at iso− seems interesting. In
the end, the results show that if the coefficient is high enough, a constant blowing AFC system can be efficient to
delay and control the massive flow separation which normally appears in the nacelle wake. The wing stall is delayed
of 1 to 2° of angle of attack and the CLmax is improved of a few percent.
21
VII. URANS simulations
The results presented in this section are from ONERA. The aerodynamic conditions are still the same. Two
URANS calculations were completed: one of the baseline configuration without flow control in stall conditions and
one with the 14-slot AFC system in pre-stall conditions. Only two computations were carried out because the CPU
cost of such unsteady cases is extremely high. The simulations took about 60 days of calculation each (CPU time). It
asked more than 6 months in real time on the ONERA HPC server. As a reminder, the grids are composed of about
70 million cells.
Figure 20 in the former section gives the average lift coefficients obtained with these two computations. They
are circled and referenced as URANS. First, the baseline configuration at 15.5° was studied to investigate the effects
of URANS on stall prediction. Indeed, it can be observed that this angle of attack corresponds to the beginning of
stall/flow separation in RANS. For the URANS computation involving an AFC system, the 14-slot device in pulsed
blowing mode (n°7) was chosen. The angle of attack of 16° was preferred because it might be considered as the
point of main interest (CLmax). Both URANS calculations have been initiated with the associated RANS solutions
(see Figure 20). In the case of the baseline configuration in stall conditions, it should be noticed that the RANS
solution presented non-negligible oscillations of the fluxes. This was not the case for the configuration with flow
control.
For the URANS computation of baseline at 15.5°, the following settings were applied:
- runs 1 to 50: time step of 1.67x10-4 s with 200 DTS sub-iterations / iteration – 50 iterations / run,
- runs 51 to 70: time step of 1.67x10-4 s with 100 DTS sub-iterations / iteration – 100 iterations / run (the
number of DTS sub-iterations was reduced to perform more iterations during a run in order to go
through the transient regime more rapidly),
- runs 71 to 100: time step of 1.67x10-3 s with 100 DTS sub-iterations / iteration – 100 iterations / run
(the time step was increased to go through the transient regime more rapidly, no problem of stability
was encountered).
These steps led to a total physical time of 5.76 seconds. The lift and drag convergence curves of this computation
can be observed in Figure 24. It can be noticed, as mentioned above, that the RANS computation (before iteration 0)
shows large oscillations due to the changing flow features at this angle of attack without control (imminent
substantial flow separation development). Then, it can be observed that the URANS simulation predicts a decreasing
lift which has been accelerated with the time step raise to finally converge to an average value 0.22 below the RANS
mean value. This decrease in CL is quite significant and leads to levels that are similar to the ones obtained with an
angle of attack of 16° in RANS, angle at which the flow separation is already massive. This is illustrated in Figure
25 which gives the RANS and URANS skin friction and friction lines at 15.5°. It confirms that the URANS at 15.5°
is close to the RANS at 16°. This can indicate that stall might appear for slightly lower angles of attack than what
the RANS predicted.
22
Figure 24: CL and CD convergence curves at 15.5°; URANS baseline (each dash is a 15h run – not all runs extracted from server).
Figure 25: Cfx distribution and friction lines at 15.5°; baseline RANS vs URANS.
The URANS simulation with the 14-slot device has needed 130 runs (10-15 hours each) on 256 cores. The
settings have been chosen according to the blowing frequency of 60 Hz (rectangular signal with duty cycle of 0.5).
A minimum of 100 time steps by signal period has been applied:
- runs 1 to 50: time step of 4.17x10-5 s with 200 DTS sub-iterations / iteration – 100 iterations / run,
- runs 51 to 90: time step of 1.67x10-4 s with 200 DTS sub-iterations / iteration – 100 iterations / run (the
time step has been slightly increased to go through the transient regime more rapidly),
- runs 91 to 125: time step of 1.67x10-4 s with 100 DTS sub-iterations / iteration – 200 iterations / run
(the number of DTS iterations has been reduced to perform more iterations during a run),
- runs 126 to 130: time step of 4.17x10-5 s with 100 DTS sub-iterations / iteration – 200 iterations / run
(the time step has been set to its initial value to investigate potential effects).
23
These steps led to a total time of 2.05 seconds. The lift and drag convergence curves of this AFC computation
are presented in Figure 26. It is important to notice that in Figure 24, for the baseline, the scale of CL axis is one
hundredth vs. one thousandth here (ten drag counts vs. one for CD). As a consequence, the variation between the 14-
slot device in constant blowing mode (RANS) and its version in pulsed blowing mode (URANS) at 16° is much
more limited than between RANS and URANS for the baseline at 15.5°. As an illustration, the CL decrease that is
observed here is only 0.035. The drag is almost not impacted. This limited lift decrease can be observed in Figure 20
with the URANS orange circle. It can be noticed that the efficiency of the 14-slot device in pulsed blowing mode at
16° is almost as high as the one of the device n°2 (constant blowing continuous slot) which yet exhibits greater
and mass flow rate values. This indicates that the alternate pulsed blowing mode is an interesting approach. Besides,
to consider the gain obtained with this AFC system computed via URANS in a consistent way when observing
Figure 20, it should be reminded that stall seem to appear earlier in URANS than in RANS for the baseline case.
Then, Figure 27 gives a direct comparison at 16° between devices n°6 (constant blowing in RANS) and n°7 (pulsed
blowing in URANS). It can be seen that both systems allow a correct control of the flow separation. The pulsed
blowing system at right exhibits flow patterns that are very similar to the ones of device n°2 at 16° in Figure 22,
which is consistent with the previous observations.
Figure 26: CL and CD convergence curves at 16°; URANS 14-slot device in pulsed blowing mode.
24
Figure 27: Cfx distribution and friction lines at 16°; 14-slot device RANS (constant) vs. URANS (pulsed).
Finally, Figure 28 shows via iso-surfaces of Mach number a time at which the slots 1, 3, 5, 7, 9, 11, and 13 are
blowing. The flow field around AFC actuators is very close to what was visible in Figure 23 except that in this case
only one slot over two is blowing which allows the mass flow rate to be cut in half without provoking a dramatic
loss in global AFC gain.
Figure 28: Iso-surfaces of Mach number (value=0.45) showing the PJA alternate injection
25
VIII. Conclusions
In the framework of the 2nd work package of the European Project AFLoNext, DLR and ONERA performed
numerical studies on a realistic high-lift aircraft configuration including slats and flap deployed for landing
conditions as well as a high bypass ratio engine. The purpose of these activities was to assess the potential of active
flow control systems to delay and contain the nacelle wake separation that eventually appears on the wing suction
side at high angles of attack.
To complete these computations, overset structured and unstructured grids were generated respectively for the
ONERA and DLR Navier-Stokes solvers. For the baseline configuration without flow control, the maximum lift
coefficient CLmax and the angle of attack at which the massive flow separation appears were determined and the flow
patterns, especially the vortex system (from pylon, slats, and strake), were studied. The DLR and ONERA results on
the baseline exhibit very good agreement. This cross-validation is satisfactory and gives good confidence in these
CFD results.
Following this work, several AFC systems were defined and evaluated, at first with RANS computations.
Different slot sizes and types (continuous vs. segmented) and/or blowing velocities were proposed. The potential of
each system was shown over a whole polar in comparison with the baseline. The gain that is obtained with an AFC
system is consistent with its coefficient value which represents the actuation force over the flow. In constant
blowing mode, all the AFC systems assessed in this study produce lift level gains after stall (typically about + 5%).
It was demonstrated that for values compatible with aircraft manufacturer requirements, a constant blowing AFC
system is efficient to delay and contain the massive flow separation which extends downstream of the engine/wing
junction without control: the wing stall can be delayed of 1 to 2° of angle of attack with the considered devices and
settings, and also be smoothed and the CLmax can be slightly increased (1 to 3%).
As a last step, URANS computations have been performed to assess the potential gain of a pulsed jet actuator.
The 14-slot system, which is probably the device the most representative of what may be tested in future AFLoNext
wind tunnel tests, was evaluated with each slot blowing out of phase with its neighbours. The baseline without
control has also been computed as a reference point in URANS. The gain obtained with the pulsed blowing system
is similar to the one of constant blowing devices that have greater and mass flow rate values, which seems to
confirm that this type of AFC approach can be promising.
26
Acknowledgments
The work described in this paper and the research leading to these results has received funding from the
European Community's Seventh Framework Programme FP7/2007-2013, under grant agreement n° 604013,
AFLoNext project. The authors thank all their partners of the work package 2.1 as well as their work package
leaders M. Meyer and J. Wild. They also thank J.L. Hantrais-Gervois (ONERA) and S. Fricke (DLR) for their
contribution. DLR thanks the Aircraft Research Association for creating the unstructured grids.
References
1.Rudnik, R., “Stall behaviour of the EUROLIFT high lift configurations,” AIAA paper, 836, 2008. 2.Geyr, H., Schade, N., van der Burg, J. W., Eliasson, P., and Esquieu, S., “CFD Prediction of Maximum Lift Effects on
Realistic High-Lift-Commercial-Aircraft-Configurations within the European project EUROLIFT II,” AIAA paper, 4299,
2007. 3.AFLoNext website: http://www.aflonext.eu/ 4.Fricke, S., Ciobaca, V., Kröhnert, A., Wild, J. and Blesbois, O., “Active Flow Control Applied at the Engine-Wing Junction,”
5th CEAS Air and Space Coneference, 2015. 5.Fricke, S., Ciobaca, V., Wild, J. and Norman, D., “Numerical Studies of Active Flow Control Applied at the Engine-Wing
Junction,” In Advances in Simulation of Wing and Nacelle Stall (pp. 397-411). Springer International Publishing, 2016. 6.Ternoy, F., Dandois, J., David, F. and Pruvost, M., “Overview of ONERA Actuators for active flow control,” AerospaceLab,
(6), 2013. 7.Greenblatt, D. and Wygnanski, I. J., “The control of flow separation by periodic excitation,” Progress in Aerospace Sciences,
36(7), 487-545, 2000. 8.Sellers, W. L., Jones, G. S. and Moore, M. D., “Flow Control Research at NASA Langley in Support of High-Lift
Augmentation,” AIAA paper 2002-6006, 2002. 9.Benoit, C., Jeanfaivre, G., Cannone, E., “Synthesis of ONERA Chimera method developed in the frame of CHANCE
program,” 31st European Rotorcraft Forum, 2005. 10.Hue, D. et al., “Validation of a near-body and off-body grid partitioning methodology for aircraft aerodynamic performance
icem-cfd-Hexa 12.Mesh Generation Software - Pointwise, http://www.pointwise.com 13.Péron, S., Benoit, C., Landier, S., and Raud, P., “Cassiopée: CFD Advanced Set of Services In an Open Python
EnvironmEnt,” 12th Symposium on Overset Grid and Solution Technology, Atlanta, 2014. 14.Leatham, M., Stokes, S., Shaw, J., Cooper, J., Appa, J., Blaylock, T., “Automatic Mesh generation for Rapid Response
Navier-Stokes Calculations,” AIAA 2000-2247, 2000. 15.Cambier, L., Heib, S., and Plot, S., “The ONERA elsA CFD Software: Input from Research and Feedback from Industry,”
Mechanics and Industry, Vol. 15(3), pp. 159-174, 2013. 16. Jameson, A., Schmidt, W., and Turkel, E., “Numerical Solution of the Euler Equations by Finite Volume Methods Using
Runge Kutta Time Stepping Schemes,” AIAA-81-1259, 1981.
27
17.Spalart, P. R., and Allmaras, S. R., "A One-Equation Turbulence Model for Aerodynamic Flows," AIAA Paper 92-0439,
1992. 18.Spalart, P. R., "Strategies for Turbulence Modeling and Simulations," International Journal of Heat and Fluid Flow, Vol. 21,
pp. 252-263, 2000. 19.Gerhold, T., “Overview of the Hybrid RANS Code TAU,” Vol. 89 of Notes on Numerical Fluid Mechanics and
Multidisciplinary Design, 2005. 20.Wada, Y., Liou, M. S., “A flux splitting scheme with high-resolution and robustness for discontinuities”, AIAA Paper 94-